J. Phys. Chem. B 2004, 108, 16077-16084
16077
Modification of ATRP Surface-Initiated Poly(hydroxyethyl methacrylate) Films with Hydrocarbon Side Chains Eric L. Brantley, Tracy C. Holmes, and G. Kane Jennings* Department of Chemical Engineering, Vanderbilt UniVersity, NashVille, Tennessee 37235 ReceiVed: June 3, 2004; In Final Form: July 26, 2004
Poly(2-hydroxyethyl methacrylate) (PHEMA) films were grown onto gold via water-accelerated, surfaceinitiated atom transfer radical polymerization (ATRP), and the resulting side chains were modified by reaction with alkanoyl chlorides (CmH2m+1COCl; m ) 1, 7, 11, 13, 15, and 17) to incorporate hydrocarbon side groups within the film. We have previously demonstrated the ability to react ∼70-80% of PHEMA hydroxyl side chains with fluorocarbon acid chlorides to prepare partially fluorinated films. Here we convert the side chains to hydrocarbon esters with diminishing conversion (80 to 40%) as m is increased and compare the resulting films with fluorocarbon-modified PHEMA. Based on IR spectra and wetting data, hydrocarbon side chains structure the film to a greater extent as m is increased. A critical chain length (m ) 15) was required to orient chains normal to the interface at the outer film surface and impart the wetting properties of a dense methyl surface. The resistances of the films against the transport of redox probes were greatly enhanced with increasing conversion of hydroxyl groups within the film and were modestly affected by film structuring to create a densely packed methyl surface. For example, capping the hydroxyl groups of PHEMA by reaction with acetyl chloride resulted in an unstructured film with >90% conversion that increased film resistance by almost four orders of magnitude over the base PHEMA film without a significant volumetric enhancement of the side chains. Also, the decreasing conversion as m is increased from 7 to 17 resulted in a decreasing film resistance even though the longer chains structured the film and surface to a greater extent. These results illustrate the important effect on barrier properties of unreacted hydroxyl groups, which possibly associate to create water and ion-diffusing pathways within the film. The barrier properties of these hydrocarbon-modified PHEMA films are comparable to those of fluorocarbon-modified PHEMA when both conversion and the molecular weight of the modifying group are similar.
Introduction In comparison with traditional processes used to prepare polymer coatings, surface-initiated polymerization methods provide numerous advantages, including chemical attachment of the polymer to the surface,1 preparation of conformal coatings on objects of various shape,2 good control over film thickness and composition,3 and ease of patterning.4,5 Much of the initial work in this area has focused on thin-film synthesis to impact materials applications, such as membrane6 or chromatographic7 separations, nanostructure modification,8,9 and etch-resistant, patterned films.4,5 In these applications, the surface and barrier properties of the polymer films are of utmost importance, and the ability to control surface and bulk film composition through molecular engineering of functional groups can have a significant impact on film performance.6,10 This manuscript examines the role of hydrocarbon side-chain length and conversion on the structural, surface, and barrier properties of surface-initiated polymer films. Hydrocarbon polymers are potentially useful in thin-films applications because they are inert, available at low cost, and resistant to penetration by moisture.11-15 However, many hydrocarbon polymers, such as polyethylene, are extremely difficult to process into ultrathin films due to solubility issues.16 A goal of our work here is to prepare polymer films that exhibit hydrocarbon-like surface and * To whom correspondence should be addressed: e-mail: jenningk@ vuse.vanderbilt.edu.
barrier properties but are actually prepared via straightforward surface-initiated strategies. Since surface and barrier properties are ultimately important in many applications, a fundamental understanding of how hydrocarbon side-chain length affects these properties would be advantageous in the molecular-level design of polymer films. We prepare these films using a two-step procedure that consists first of water-accelerated, surface-initiated atom transfer radical polymerization (ATRP)3,9,17-21 followed by reaction of the grown film with a hydrocarbon species to introduce functionality. The advantages of this approach are that the polymer film is covalently attached to an initiator molecule on the surface, a tremendous amount of control of film growth kinetics over a wide thickness range is achieved by performing ATRP in water,3,19 and film composition can be tailored by simple reaction to add appropriate molecular groups.3,6,10,21 This methodology allows us to use a base film as a test stage for creation of numerous polymer films with varying chemical composition. The surface-initiated growth of polymer films via ATRP typically exhibits sluggish kinetics that limit film thickness,1,22,23 but the solution polymerization of a hydrophilic, water-soluble monomer may be accelerated by using water as solvent.17,18 For the surface-initiated case, rather thick films (>100 nm) can now be grown in a controlled fashion using water-accelerated ATRP.3,19 However, since acrylates or methacrylates having long alkyl side chains are not hydrophilic, the surface-initiated ATRP
10.1021/jp0476038 CCC: $27.50 © 2004 American Chemical Society Published on Web 09/16/2004
16078 J. Phys. Chem. B, Vol. 108, No. 41, 2004 SCHEME 1: Acylation of PHEMA with Hydrocarbon Acid Chlorides
of such monomers would not likely produce films that are sufficiently thick to be effective barriers.24 To achieve much thicker hydrocarbon films, we first polymerize the water-soluble, hydrophilic monomer hydroxyethyl methacrylate (HEMA) via water-accelerated, surface-initiated ATRP.3,10,19 This method results in the controlled growth of moderately thick (>100 nm) PHEMA films. To gain even more control over film characteristics and to produce hydrocarbon-rich films, the hydroxylterminated side chains throughout the PHEMA film are reacted to introduce desirable chemical functionality. Huang et al.3 previously reacted PHEMA with short hydrocarbon acid chlorides, thus demonstrating the ability to introduce groups into the polymer film via a simple acylation reaction of the side chains. Other recent work has exploited the hydroxyl groups of PHEMA by reaction with trimethylchlorosilane to make the film more hydrophobic and improve its etching resistance.5 Additionally, modification of hydroxyl groups with C15H31COCl has been shown to dramatically improve surface hydrophobicity of thin (e40 nm) surface-initiated polyglycidol brushes grown via anionic ring-opening polymerization on silicon.25 We have previously investigated the reaction of hydroxyl groups of PHEMA with fluorocarbon acid chlorides and extensively analyzed the structure and composition of the films as well as the surface and barrier properties.10 We observed that a longer fluorocarbon side chain (of length 7) imparted structure to the film, both at the surface where the chains align along the normal at the air-film interface and also in the bulk where the chains pack to maximize interchain interactions, whereas a shorter fluorocarbon chain (of length 3) offered little or no structuring. The current work conducts the same modification but uses hydrocarbon acid chlorides (CH3(CH2)m-1COCl; m ) 1, 7, 11, 13, 15, and 17; Scheme 1), allowing the comparison of film properties as a function of side-group composition, chain length, and molecular weight. Analogous to the work with fluorocarbon side chains, we identify hydrocarbon chain lengths that give structure to the polymer films at the surface and in the bulk. Other work has been done on films containing long hydrocarbon chains. Kraft and Moore26 have investigated the effects of reacting PHEMA microgels with hydrocarbon acid chlorides (m ranging from 1 to 15) to form fatty acid layers that delay microgel expansion. They found that acetyl-modified and unmodified microgels expand readily, that rather short (m ) 4)
Brantley et al. alkyl chains offer maximum resistance to microgel expansion, and that as the chain length is further increased the resistance decreases. Varying alkyl side chain length has also been tested on Langmuir-Blodgett films, specifically investigating the role of chain length on the structure and order of the film packing. Of particular relevance, acylethylenimines having short hydrocarbon side chains were found to exist in an easily compressible, liquid-like state.27 As the side-chain length was increased, the monolayers became more condensed, until at chain lengths greater than 13, the resulting monolayers became crystallized, rigid structures. Similarly, a hydrocarbon side chain length of 18 in polyimide Langmuir-Blodgett films resulted in crystallized structures, as verified by IR peak positions for methylene stretching modes.28 As for surface-initiated polymers, Sto¨hr and Ru¨he29 grew various n-alkyl methacrylates from physisorbed poly(caprolactone) macroinitiators on silicon oxide. They found that the surface became increasingly hydrophobic as the length of the polymer side chain was increased from a methyl to a stearyl group. However, to our knowledge, the combined structural, surface, and barrier properties of polymer films with varying alkyl side chain length have not been investigated. Our general film preparation methods allow us to make these film property comparisons among various hydrocarbon-modified films produced in this work along with comparisons to our previously reported fluorocarbon-modified PHEMA films. Experimental Section Materials. CuCl (99.995+%), CuBr2 (99.999%), 2,2′-bipyridine (bpy, 99+%), 2-hydroxyethyl methacrylate (HEMA, >99%), pyridine (99+%), lauroyl chloride (C11H23COCl, 98%), myristoyl chloride (C13H27COCl, 97%), palmitoyl chloride (C15H31COCl, 98%), K3Fe(CN)6 (99+%), K4Fe(CN)6‚3H2O (99%), and hexadecane (99%) were used as received from Aldrich. N,N-Dimethylformamide (DMF, 99.9%), isooctane (99%), acetyl chloride (CH3COCl, 98%), octanoyl chloride (C7H15COCl, 99%), stearoyl chloride (C17H35COCl, >99%), and Na2SO4 (anhydrous) were used as received from Fisher. Gold shot (99.99%) and chromium-coated tungsten filaments were obtained from J&J Materials and R. D. Mathis, respectively. Silicon (100) wafers (Montco Silicon) were rinsed with ethanol and deionized water and dried with nitrogen. Ethanol (AAPER, absolute) was used as received. Deionized water (16.7 MΩ‚ cm) was purified with a Modu-Pure system and used as a solvent during polymerization and for rinsing. An initiator-terminated disulfide, (BrC(CH3)2COO(CH2)11S)2, was synthesized as described in the literature.4 Preparation of Gold Substrates. Gold substrates were prepared by evaporating chromium (100 Å) and gold (1250 Å) in sequence onto silicon (100) wafers at rates of 1-2 Å s-1 in a diffusion-pumped chamber with a base pressure of 4 × 10-6 Torr. After removal from the evaporation chamber, the wafers were typically cut into 1 cm × 3 cm pieces. Polymerization. Gold substrates were first placed in a 1 mM ethanol solution of (BrC(CH3)2COO(CH2)11S)2 for 24 h. The initiated samples were then rinsed with ethanol, dried with nitrogen, and placed in vials that were subsequently degassed and backfilled with nitrogen. A CuI/CuII/bpy (69 mM CuCl, 20 mM CuBr2, 195 mM bpy) system in a 50:50 v:v water/HEMA solution was used for polymerization.3 The mixture was placed in a Schlenk flask sealed with a rubber septum and was degassed by performing three freeze-pump-thaw cycles. This was followed by transfer of the solution via cannula into vials containing up to six samples each. After polymerizing for 12 h at room temperature, the samples were thoroughly rinsed with
Poly(Hydroxyethyl Methacrylate) Films water and DMF and then dried with nitrogen. As measured by ellipsometry with samples from three different batches, PHEMA film thicknesses were 217 ( 11 nm under these conditions. Film Acylation. Gold surfaces with PHEMA films were exposed to 20 mM solutions of acetyl chloride (CH3COCl), octanoyl chloride (C7H15COCl), lauroyl chloride (C11H23COCl), myristoyl chloride (C13H27COCl), palmitoyl chloride (C15H31COCl), or stearoyl chloride (C17H35COCl) with 25 mM pyridine in isooctane for 24 h to give PHEMA films with acylated side chains (Scheme 1). The concentration of 20 mM is near or above the solubility limit of the longer-chain acid chlorides in isooctane. After reaction, the films were rinsed with isooctane and ethanol and dried with nitrogen. The various alkyl-modified PHEMA films are referred to by the length of the hydrocarbon chain (i.e. C7H15COCl-modified PHEMA is H7). Characterization Methods. Polymer film properties were evaluated using the following methods before acylation to provide a baseline measurement and once again after acylation to track changes in film properties. Reflectance absorption infrared spectroscopy (RAIRS) was performed using a Bio-Rad Excalibur FTS-3000 infrared spectrometer. The p-polarized light was incident at 80° from the surface normal. The instrument was run in single reflection mode and equipped with a Universal sampling accessory. A liquid-nitrogen-cooled, narrow-band MCT detector was used to detect reflected light. Spectral resolution was 2 cm-1 after triangular apodization. Each spectrum was accumulated over 1000 scans using a deuterated octadecanethiol-d37 self-assembled monolayer on gold as the background. Ellipsometry measurements were taken on a J. A. Woollam Co. M-2000DI variable angle spectroscopic ellipsometer with WVASE32 software for modeling. Measurements at three spots per sample were taken with light incident at a 75° angle from the surface normal using wavelengths from 400 to 800 nm. Optical constants for a bare gold substrate, cut from the same wafer as the samples to be characterized, were measured by ellipsometry and used as the baseline for all polymer film samples. Film thickness of the polymer layer on samples, regardless of modification, was obtained using a two-term Cauchy layer model, allowing the modeling software to fit thickness as well as the two Cauchy terms defining the refractive index, n. Typical values for n from the modeling software ranged between 1.45 and 1.50. Film thicknesses reported herein are generally lower but believed to be more accurate than in our previous work10 due to changes in our ellipsometry procedures. A Rame-Hart contact angle goniometer with a microliter syringe was used to measure advancing and receding contact angles on static drops of water and hexadecane on the polymer surfaces. The needle tip of the syringe remained inside the liquid drop while measurements were taken on both sides of ∼5 µL drops. Reported values and ranges represent the average and standard deviation of values obtained from at least five independent sample preparations. Electrochemical impedance spectroscopy (EIS) was performed with a Gamry Instruments CMS300 impedance system interfaced to a personal computer. A flat-cell (EG&G Instruments) was used to expose only 1 cm2 of each sample to an aqueous solution containing electrolyte and redox probes while preventing sample edges from being exposed. The electrochemical cell consisted of an aqueous solution of 1 mM K3Fe(CN)6, 1 mM K4Fe(CN)6‚3H2O, and 0.1 M Na2SO4 with a Ag/AgCl/ saturated KCl reference electrode, a gold substrate counter electrode, and a gold substrate containing the film to be studied as the working electrode. All data were collected in the range
J. Phys. Chem. B, Vol. 108, No. 41, 2004 16079
Figure 1. Reflectance absorption IR spectra of PHEMA and a representative hydrocarbon-modified PHEMA sample, H7. Regions of interest are labeled on the plot.
TABLE 1: Position (cm-1) of Selected IR Peaks and Average Conversion for Modified PHEMA Films film
ν (CdO)a
H1 H7 H11 H13 H15 H17 H17/H1
1750 1744 1744 1742 1742 1737 1750
νs (CH2) 2858 2855 2855 2852 2852 2852
νa (CH2)
χ (%)
2930 2926 2925 2922 2921 2920
93 ( 3 82 ( 5 77 ( 6 68 ( 5 64 ( 4 37 ( 6 82 ( 4
a These positions are typical for each film but can shift slightly as conversion deviates from the average values.
from 10-1 to 104 Hz using 10 points per decade and were fit with an appropriate equivalent circuit model (vide infra) to determine resistance and capacitance values. Reported values and ranges for resistance and capacitance represent the average and standard deviation of values obtained from at least four independent sample preparations. Results and Discussion Film Composition and Structure. We used RAIRS to monitor compositional changes within the polymer films due to acylation by the various acid chlorides. Figure 1 shows a survey IR spectrum for a representative hydrocarbon-modified PHEMA film, H7, along with the spectrum for PHEMA. All other films exhibit the same peaks in the IR spectrum as H7 but at slightly different positions and intensities. The most readily quantifiable change in the spectra that is indicative of successful acylation is diminution of the hydroxyl peak in the region from 3100 to 3700 cm-1, as the hydroxyl side chains of PHEMA are converted to esters. We have used the diminution of integrated hydroxyl peak area (AOH) to estimate conversion (χ) of the hydroxyl side chains as
χ)1-
AOH, acylated PHEMA AOH, PHEMA
(1)
which assumes that any change in integrated hydroxyl peak area is due solely to acylation and not to orientational changes of unreacted hydroxyl groups. Table 1 lists conversions, calculated using Eq 1, from at least five independent preparations for each acid chloride. As the chain length of the hydrocarbon side chain is increased, conversion drops substantially, suggesting that molecular size limits penetration and subsequent reaction of acid chlorides within the film. For the modification with the longest side group (H17), conversion was significantly lower than that for the other hydrocarbon films. Several attempts were made, without success, to increase conversion for H17 by using
16080 J. Phys. Chem. B, Vol. 108, No. 41, 2004
Figure 2. Carbonyl stretching region of IR spectra for PHEMA and hydrocarbon-modified PHEMA films. The dashed line is at 1733 cm-1, the position of the carbonyl peak for PHEMA.
different solvents and raising temperature. For completeness, results for H1 and a copolymer film, H17/H1, are included in tables along with the other films. These films are not discussed until later in this section. The acylation reaction studied here causes two other differences in the IR spectra of modified PHEMA films as compared to PHEMA: introduction of an additional ester linkage and the presence of hydrocarbon chains of varying length on each modified side chain. The hydrocarbon ester from acylation of the side chain results in a second carbonyl peak in the spectrum around 1750 cm-1 and enhanced C-O stretching peaks from 1100 to 1300 cm-1. The introduction of long hydrocarbon chains into the polymer film results in a significant increase in the intensity of C-H bending (1400 to 1500 cm-1) and C-H stretching (2800 to 3000 cm-1) modes in the IR spectrum. The carbonyl peak resulting from acylation is not easily distinguished from the original PHEMA ester carbonyl peak (Figure 2) due to a similar chemical environment and therefore similar peak positions. The peak for the carbonyl associated with the PHEMA ester appears at 1733 cm-1, whereas the carbonyl peak due to acylation appears at somewhat higher wavenumbers (∼1750 cm-1, as evidenced by this peak position for H1) but still relatively close to the PHEMA carbonyl peak. Upon acylation, one broader peak appears in the region, representing the combined intensities of the PHEMA carbonyl and the carbonyl from the acylation reaction. From Figure 2 and Table 1, the resultant carbonyl peak position shifts with increasing hydrocarbon chain length (m). This shift is due to lower conversion within the film as m increases, meaning that less IR peak intensity is introduced at higher wavenumbers by the new carbonyl. Furthermore, the overall intensity of the combined carbonyl peak is also reduced as m is increased because of diminishing conversion. Addition of alkyl groups to the PHEMA side chains also results in a significant increase in C-H bending and stretching bands in the IR spectrum, and the position of these peaks gives an indication of chain packing and crystallinity. As compared to the spectrum for PHEMA (Figure 1), those for modified PHEMA show a sharper CH2 bending peak at ∼1470 cm-1. However, there is no peak splitting as observed in crystalline polymethylene lattices with orthorhombic chain packing.16 Symmetric and asymmetric methylene stretching modes, νs(CH2) and νa(CH2) respectively, become much more prominent and sharper after addition of alkyl side chains (Figure 3). From previous IR studies of polymethylene chains,16,30 a νs(CH2) mode positioned at ∼2850 cm-1 and a νa(CH2) mode positioned at
Brantley et al.
Figure 3. C-H stretching region of IR spectra for PHEMA and hydrocarbon-modified PHEMA films. The dashed lines are positioned at 2852 and 2921 cm-1, corresponding to the lowest peak positions observed for νs(CH2) and νa(CH2) in these films.
∼2918 cm-1 are indicative of highly crystalline chain packing while shifts to higher wavenumbers indicate less dense, more liquid-like packing.31 Figure 3 shows that increasing the hydrocarbon chain length results in more crystalline chain packing (fewer gauche conformers) within the film as the νa(CH2) position decreases from 2930 (m ) 7) to 2921 (m ) 17) cm-1. The improved crystallinity observed for longer alkyl side groups is consistent with greater van der Waals interactions among hydrocarbon chains, which increase proportionally with chain length and facilitate chain structuring within the film.32 Additionally, in copolymer systems having one component that packs efficiently and one that packs poorly (i.e. ethylene-butene copolymers33), the efficient-packing chains tend to order while the poorer-packing chains do not. In our system, increasing hydrocarbon chain length results in more crystalline packing because the van der Waals interactions between the alkyl chains become increasingly important as compared with the interactions between other groups within the polymer film. Film Thickness. Table 2 shows film thicknesses obtained by ellipsometry before and after acylation with hydrocarbon acid chlorides. Starting with a PHEMA film, the polymer layer expands upon acylation with hydrocarbon acid chlorides to accommodate the additional side-chain volume. The observed trends roughly agree (within ∼30%) with theoretical estimates based on conversion and molecular weight considerations, which have been used previously3 to model the addition of short hydrocarbon groups to PHEMA. With the exception of H17, which exhibits exceptionally low conversion, the addition of long hydrocarbon groups (m ) 7 to 15) resulted in slightly more than a doubling of polymer film thickness. Since the hydrocarbon chains exhibit lower conversions from H7 to H15, the film thickness change remains relatively constant for this series. Surface Wetting Properties. Advancing and receding contact angles (θA and θR) of water were measured for all polymer films to provide a measure of surface hydrophobicity. Table 3 gives average contact angles for the modified films along with the values for PHEMA and the bromine-terminated initiator monolayer for comparison. Both the advancing contact angle and the contact angle hysteresis (θA - θR) of water on the initiator are indicative of a smooth, bromine-terminated monolayer.34 Once PHEMA is grown from the initiator, the advancing contact angle is not significantly altered, but the hysteresis becomes quite high, suggesting a rough and/or chemically heterogeneous surface. Hydrocarbon-modified PHEMA films (m ) 7-17) show much higher advancing water contact angles, ranging from 104° to
Poly(Hydroxyethyl Methacrylate) Films
J. Phys. Chem. B, Vol. 108, No. 41, 2004 16081
TABLE 2: Thickness Change of PHEMA Films upon Acylation with Hydrocarbon Acid Chlorides thickness (nm)
a
film
before
after
observed increase
repeat unit MW (g mol-1)
theoretical increase (MW)
H1 H7 H11 H13 H15 H17 H17/H1
216 215 224 205 231 231 208
295 452 508 458 489 322 318
37% 110% 127% 124% 111% 40% 53%
172 256 312 341 369 397 -
30% 81% 116% 121% 119% 75% 87%a
Based on 35% conversion for H17 and 47% for H1.
TABLE 3: Water and Hexadecane Advancing and Receding Contact Angles (°) for Films on Gold water
hexadecane
film
θA
θR
θA
θR
initiator PHEMA H1 H7 H11 H13 H15 H17
80 ( 2 75 ( 3 70 ( 2 104 ( 2 113 ( 1 117 ( 1 117 ( 2 115 ( 3
74 ( 2 23 ( 2 52 ( 3 60 ( 4 59 ( 4 70 ( 4 65 ( 5 77 ( 7
